Torch Geometric: Node Classification on Large Graphs

Hi all,

I am a newbee to the GNN world. I wanted to run some node classification over my graph. The issue is that the graph itself doesn’t fit on one GPU. Is there a way that we can run the classification on subgraphs so instead of loading the entire graph on GPU, I process a subgraph at each learning iteration? Something like batch-wise processing of images.

Thanks.